refactor engine's search language handling

Add match_language function in utils to match any user given
language code with a list of engine's supported languages.

Also add language_aliases dict on each engine to translate
standard language codes into the custom codes used by the engine.

searx/engines/__init__.py

@@ -20,13 +20,14 @@ import sys
 import threading
 from os.path import realpath, dirname
 from io import open
+from babel.localedata import locale_identifiers
 from flask_babel import gettext
 from operator import itemgetter
 from json import loads
 from requests import get
 from searx import settings
 from searx import logger
-from searx.utils import load_module
+from searx.utils import load_module, match_language
 
 
 logger = logger.getChild('engines')
@@ -38,6 +39,8 @@ engines = {}
 categories = {'general': []}
 
 languages = loads(open(engine_dir + '/../data/engines_languages.json', 'r', encoding='utf-8').read())
+babel_langs = [lang_parts[0] + '-' + lang_parts[-1] if len(lang_parts) > 1 else lang_parts[0]
+               for lang_parts in (lang_code.split('_') for lang_code in locale_identifiers())]
 
 engine_shortcuts = {}
 engine_default_args = {'paging': False,
@@ -97,6 +100,22 @@ def load_engine(engine_data):
     if engine_data['name'] in languages:
         setattr(engine, 'supported_languages', languages[engine_data['name']])
 
+    # find custom aliases for non standard language codes
+    if hasattr(engine, 'supported_languages'):
+        if hasattr(engine, 'language_aliases'):
+            language_aliases = getattr(engine, 'language_aliases')
+        else:
+            language_aliases = {}
+
+        for engine_lang in getattr(engine, 'supported_languages'):
+            iso_lang = match_language(engine_lang, babel_langs, fallback=None)
+            if iso_lang and iso_lang != engine_lang and not engine_lang.startswith(iso_lang) and \
+               iso_lang not in getattr(engine, 'supported_languages'):
+                language_aliases[iso_lang] = engine_lang
+
+        if language_aliases:
+            setattr(engine, 'language_aliases', language_aliases)
+
     # assign language fetching method if auxiliary method exists
     if hasattr(engine, '_fetch_supported_languages'):
         setattr(engine, 'fetch_supported_languages',

searx/engines/archlinux.py

@@ -99,13 +99,13 @@ supported_languages = dict(lang_urls, **main_langs)
 
 # do search-request
 def request(query, params):
-    # translate the locale (e.g. 'en_US') to language code ('en')
+    # translate the locale (e.g. 'en-US') to language code ('en')
     language = locale_to_lang_code(params['language'])
 
     # if our language is hosted on the main site, we need to add its name
     # to the query in order to narrow the results to that language
     if language in main_langs:
-        query += '(' + main_langs[language] + ')'
+        query += b' (' + main_langs[language] + b')'
 
     # prepare the request parameters
     query = urlencode({'search': query})

searx/engines/bing.py

@@ -16,12 +16,14 @@
 from lxml import html
 from searx.engines.xpath import extract_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
 paging = True
 language_support = True
 supported_languages_url = 'https://www.bing.com/account/general'
+language_aliases = {'zh-CN': 'zh-CHS', 'zh-TW': 'zh-CHT', 'zh-HK': 'zh-CHT'}
 
 # search-url
 base_url = 'https://www.bing.com/'
@@ -32,9 +34,9 @@ search_string = 'search?{query}&first={offset}'
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
 
-    lang = params['language'].split('-')[0].upper()
+    lang = match_language(params['language'], supported_languages, language_aliases)
 
-    query = u'language:{} {}'.format(lang, query.decode('utf-8')).encode('utf-8')
+    query = u'language:{} {}'.format(lang.split('-')[0].upper(), query.decode('utf-8')).encode('utf-8')
 
     search_path = search_string.format(
         query=urlencode({'q': query}),

searx/engines/bing_images.py

@@ -19,6 +19,7 @@ from lxml import html
 from json import loads
 import re
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['images']
@@ -46,26 +47,6 @@ safesearch_types = {2: 'STRICT',
 _quote_keys_regex = re.compile('({|,)([a-z][a-z0-9]*):(")', re.I | re.U)
 
 
-# get supported region code
-def get_region_code(lang, lang_list=None):
-    region = None
-    if lang in (lang_list or supported_languages):
-        region = lang
-    elif lang.startswith('no'):
-        region = 'nb-NO'
-    else:
-        # try to get a supported country code with language
-        lang = lang.split('-')[0]
-        for lc in (lang_list or supported_languages):
-            if lang == lc.split('-')[0]:
-                region = lc
-                break
-    if region:
-        return region.lower()
-    else:
-        return 'en-us'
-
-
 # do search-request
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
@@ -74,7 +55,7 @@ def request(query, params):
         query=urlencode({'q': query}),
         offset=offset)
 
-    language = get_region_code(params['language'])
+    language = match_language(params['language'], supported_languages).lower()
 
     params['cookies']['SRCHHPGUSR'] = \
         'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')

searx/engines/bing_news.py

@@ -14,8 +14,8 @@
 from datetime import datetime
 from dateutil import parser
 from lxml import etree
-from searx.utils import list_get
-from searx.engines.bing import _fetch_supported_languages, supported_languages_url
+from searx.utils import list_get, match_language
+from searx.engines.bing import _fetch_supported_languages, supported_languages_url, language_aliases
 from searx.url_utils import urlencode, urlparse, parse_qsl
 
 # engine dependent config
@@ -71,7 +71,7 @@ def request(query, params):
 
     offset = (params['pageno'] - 1) * 10 + 1
 
-    language = params['language']
+    language = match_language(params['language'], supported_languages, language_aliases)
 
     params['url'] = _get_url(query, language, offset, params['time_range'])
 

searx/engines/bing_videos.py

@@ -12,9 +12,10 @@
 
 from json import loads
 from lxml import html
-from searx.engines.bing_images import _fetch_supported_languages, supported_languages_url, get_region_code
+from searx.engines.bing_images import _fetch_supported_languages, supported_languages_url
 from searx.engines.xpath import extract_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 
 categories = ['videos']
@@ -47,8 +48,8 @@ def request(query, params):
         'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')
 
     # language cookie
-    region = get_region_code(params['language'], lang_list=supported_languages)
-    params['cookies']['_EDGE_S'] = 'mkt=' + region + '&F=1'
+    language = match_language(params['language'], supported_languages).lower()
+    params['cookies']['_EDGE_S'] = 'mkt=' + language + '&F=1'
 
     # query and paging
     params['url'] = search_url.format(query=urlencode({'q': query}),

searx/engines/dailymotion.py

@@ -15,6 +15,7 @@
 from json import loads
 from datetime import datetime
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['videos']
@@ -32,7 +33,7 @@ supported_languages_url = 'https://api.dailymotion.com/languages'
 
 # do search-request
 def request(query, params):
-    locale = params['language']
+    locale = match_language(params['language'], supported_languages)
 
     params['url'] = search_url.format(
         query=urlencode({'search': query, 'localization': locale}),

searx/engines/duckduckgo.py

@@ -18,6 +18,7 @@ from json import loads
 from searx.engines.xpath import extract_text
 from searx.poolrequests import get
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
@@ -26,6 +27,16 @@ language_support = True
 supported_languages_url = 'https://duckduckgo.com/util/u172.js'
 time_range_support = True
 
+language_aliases = {
+    'ar-SA': 'ar-XA',
+    'es-419': 'es-XL',
+    'ja': 'jp-JP',
+    'ko': 'kr-KR',
+    'sl-SI': 'sl-SL',
+    'zh-TW': 'tzh-TW',
+    'zh-HK': 'tzh-HK'
+}
+
 # search-url
 url = 'https://duckduckgo.com/html?{query}&s={offset}&dc={dc_param}'
 time_range_url = '&df={range}'
@@ -42,34 +53,12 @@ content_xpath = './/a[@class="result__snippet"]'
 
 
 # match query's language to a region code that duckduckgo will accept
-def get_region_code(lang, lang_list=None):
-    # custom fixes for languages
-    if lang[:2] == 'ja':
-        region_code = 'jp-jp'
-    elif lang[:2] == 'sl':
-        region_code = 'sl-sl'
-    elif lang == 'zh-TW':
-        region_code = 'tw-tzh'
-    elif lang == 'zh-HK':
-        region_code = 'hk-tzh'
-    elif lang[-2:] == 'SA':
-        region_code = 'xa-' + lang.split('-')[0]
-    elif lang[-2:] == 'GB':
-        region_code = 'uk-' + lang.split('-')[0]
-    else:
-        region_code = lang.split('-')
-        if len(region_code) == 2:
-            # country code goes first
-            region_code = region_code[1].lower() + '-' + region_code[0].lower()
-        else:
-            # tries to get a country code from language
-            region_code = region_code[0].lower()
-            for lc in (lang_list or supported_languages):
-                lc = lc.split('-')
-                if region_code == lc[0]:
-                    region_code = lc[1].lower() + '-' + lc[0].lower()
-                    break
-    return region_code
+def get_region_code(lang, lang_list=[]):
+    lang_code = match_language(lang, lang_list, language_aliases, 'wt-WT')
+    lang_parts = lang_code.split('-')
+
+    # country code goes first
+    return lang_parts[1].lower() + '-' + lang_parts[0].lower()
 
 
 # do search-request
@@ -79,7 +68,7 @@ def request(query, params):
 
     offset = (params['pageno'] - 1) * 30
 
-    region_code = get_region_code(params['language'])
+    region_code = get_region_code(params['language'], supported_languages)
     params['url'] = url.format(
         query=urlencode({'q': query, 'kl': region_code}), offset=offset, dc_param=offset)
 

searx/engines/duckduckgo_definitions.py

@@ -2,9 +2,9 @@ import json
 from lxml import html
 from re import compile
 from searx.engines.xpath import extract_text
-from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url
+from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url, language_aliases
 from searx.url_utils import urlencode
-from searx.utils import html_to_text
+from searx.utils import html_to_text, match_language
 
 url = 'https://api.duckduckgo.com/'\
     + '?{query}&format=json&pretty=0&no_redirect=1&d=1'
@@ -24,7 +24,8 @@ def result_to_text(url, text, htmlResult):
 
 def request(query, params):
     params['url'] = url.format(query=urlencode({'q': query}))
-    params['headers']['Accept-Language'] = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages, language_aliases)
+    params['headers']['Accept-Language'] = language.split('-')[0]
     return params
 
 

searx/engines/duckduckgo_images.py

@@ -15,7 +15,10 @@
 
 from json import loads
 from searx.engines.xpath import extract_text
-from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url, get_region_code
+from searx.engines.duckduckgo import (
+    _fetch_supported_languages, supported_languages_url,
+    get_region_code, language_aliases
+)
 from searx.poolrequests import get
 from searx.url_utils import urlencode
 

searx/engines/google.py

@@ -14,6 +14,7 @@ from lxml import html, etree
 from searx.engines.xpath import extract_text, extract_url
 from searx import logger
 from searx.url_utils import urlencode, urlparse, parse_qsl
+from searx.utils import match_language
 
 logger = logger.getChild('google engine')
 
@@ -165,22 +166,20 @@ def extract_text_from_dom(result, xpath):
 def request(query, params):
     offset = (params['pageno'] - 1) * 10
 
-    # temporary fix until a way of supporting en-US is found
-    if params['language'] == 'en-US':
-        params['language'] = 'en-GB'
-
-    if params['language'][:2] == 'jv':
-        language = 'jw'
-        country = 'ID'
-        url_lang = 'lang_jw'
+    language = match_language(params['language'], supported_languages)
+    language_array = language.split('-')
+    if params['language'].find('-') > 0:
+        country = params['language'].split('-')[1]
+    elif len(language_array) == 2:
+        country = language_array[1]
     else:
-        language_array = params['language'].lower().split('-')
-        if len(language_array) == 2:
-            country = language_array[1]
-        else:
-            country = 'US'
-        language = language_array[0] + ',' + language_array[0] + '-' + country
-        url_lang = 'lang_' + language_array[0]
+        country = 'US'
+
+    # temporary fix until a way of supporting en-US is found
+    if language == 'en-US':
+        country = 'GB'
+
+    url_lang = 'lang_' + language
 
     if use_locale_domain:
         google_hostname = country_to_hostname.get(country.upper(), default_hostname)
@@ -196,7 +195,7 @@ def request(query, params):
     if params['time_range'] in time_range_dict:
         params['url'] += time_range_search.format(range=time_range_dict[params['time_range']])
 
-    params['headers']['Accept-Language'] = language
+    params['headers']['Accept-Language'] = language + ',' + language + '-' + country
     params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8'
 
     params['google_hostname'] = google_hostname

searx/engines/google_news.py

@@ -13,6 +13,7 @@
 from lxml import html
 from searx.engines.google import _fetch_supported_languages, supported_languages_url
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # search-url
 categories = ['news']
@@ -50,8 +51,9 @@ def request(query, params):
     params['url'] = search_url.format(query=urlencode({'q': query}),
                                       search_options=urlencode(search_options))
 
-    language_array = params['language'].lower().split('-')
-    params['url'] += '&lr=lang_' + language_array[0]
+    language = match_language(params['language'], supported_languages).split('-')[0]
+    if language:
+        params['url'] += '&lr=lang_' + language
 
     return params
 

searx/engines/qwant.py

@@ -14,6 +14,7 @@ from datetime import datetime
 from json import loads
 from searx.utils import html_to_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = None
@@ -45,16 +46,8 @@ def request(query, params):
                                    offset=offset)
 
     # add language tag
-    if params['language'] == 'no' or params['language'].startswith('no-'):
-        params['language'] = params['language'].replace('no', 'nb', 1)
-    if params['language'].find('-') < 0:
-        # tries to get a country code from language
-        for lang in supported_languages:
-            lc = lang.split('-')
-            if params['language'] == lc[0]:
-                params['language'] = lang
-                break
-    params['url'] += '&locale=' + params['language'].replace('-', '_').lower()
+    language = match_language(params['language'], supported_languages)
+    params['url'] += '&locale=' + language.replace('-', '_').lower()
 
     return params
 

searx/engines/swisscows.py

@@ -14,6 +14,7 @@ from json import loads
 import re
 from lxml.html import fromstring
 from searx.url_utils import unquote, urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general', 'images']
@@ -35,11 +36,8 @@ regex_img_url_remove_start = re.compile(b'^https?://i\.swisscows\.ch/\?link=')
 
 # do search-request
 def request(query, params):
-    if params['language'].split('-')[0] == 'no':
-        region = 'nb-NO'
-    else:
-        region = params['language']
-        ui_language = params['language'].split('-')[0]
+    region = match_language(params['language'], supported_languages)
+    ui_language = region.split('-')[0]
 
     search_path = search_string.format(
         query=urlencode({'query': query, 'uiLanguage': ui_language, 'region': region}),

searx/engines/wikidata.py

@@ -16,6 +16,7 @@ from searx.poolrequests import get
 from searx.engines.xpath import extract_text
 from searx.engines.wikipedia import _fetch_supported_languages, supported_languages_url
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 from json import loads
 from lxml.html import fromstring
@@ -56,7 +57,7 @@ calendar_name_xpath = './/sup[contains(@class,"wb-calendar-name")]'
 
 
 def request(query, params):
-    language = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages).split('-')[0]
 
     params['url'] = url_search.format(
         query=urlencode({'label': query, 'language': language}))
@@ -68,7 +69,7 @@ def response(resp):
     html = fromstring(resp.text)
     wikidata_ids = html.xpath(wikidata_ids_xpath)
 
-    language = resp.search_params['language'].split('-')[0]
+    language = match_language(resp.search_params['language'], supported_languages).split('-')[0]
 
     # TODO: make requests asynchronous to avoid timeout when result_count > 1
     for wikidata_id in wikidata_ids[:result_count]:

searx/engines/wikipedia.py

@@ -13,6 +13,7 @@
 from json import loads
 from lxml.html import fromstring
 from searx.url_utils import quote, urlencode
+from searx.utils import match_language
 
 # search-url
 base_url = u'https://{language}.wikipedia.org/'
@@ -30,13 +31,7 @@ supported_languages_url = 'https://meta.wikimedia.org/wiki/List_of_Wikipedias'
 
 # set language in base_url
 def url_lang(lang):
-    lang = lang.split('-')[0]
-    if lang not in supported_languages:
-        language = 'en'
-    else:
-        language = lang
-
-    return language
+    return match_language(lang, supported_languages).split('-')[0]
 
 
 # do search-request

searx/engines/yahoo.py

@@ -14,6 +14,7 @@
 from lxml import html
 from searx.engines.xpath import extract_text, extract_url
 from searx.url_utils import unquote, urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
@@ -39,6 +40,8 @@ time_range_dict = {'day': ['1d', 'd'],
                    'week': ['1w', 'w'],
                    'month': ['1m', 'm']}
 
+language_aliases = {'zh-CN': 'zh-CHS', 'zh-TW': 'zh-CHT', 'zh-HK': 'zh-CHT'}
+
 
 # remove yahoo-specific tracking-url
 def parse_url(url_string):
@@ -70,23 +73,16 @@ def _get_url(query, offset, language, time_range):
                                         lang=language)
 
 
-def _get_language(params):
-    if params['language'][:2] == 'zh':
-        if params['language'] == 'zh' or params['language'] == 'zh-CH':
-            return 'szh'
-        else:
-            return 'tzh'
-    else:
-        return params['language'].split('-')[0]
-
-
 # do search-request
 def request(query, params):
     if params['time_range'] and params['time_range'] not in time_range_dict:
         return params
 
     offset = (params['pageno'] - 1) * 10 + 1
-    language = _get_language(params)
+    language = match_language(params['language'], supported_languages, language_aliases)
+    if language not in language_aliases.values():
+        language = language.split('-')[0]
+    language = language.replace('-', '_').lower()
 
     params['url'] = _get_url(query, offset, language, params['time_range'])
 
@@ -145,7 +141,11 @@ def _fetch_supported_languages(resp):
     dom = html.fromstring(resp.text)
     options = dom.xpath('//div[@id="yschlang"]/span/label/input')
     for option in options:
-        code = option.xpath('./@value')[0][5:].replace('_', '-')
+        code_parts = option.xpath('./@value')[0][5:].split('_')
+        if len(code_parts) == 2:
+            code = code_parts[0] + '-' + code_parts[1].upper()
+        else:
+            code = code_parts[0]
         supported_languages.append(code)
 
     return supported_languages

searx/engines/yahoo_news.py

@@ -13,9 +13,12 @@ import re
 from datetime import datetime, timedelta
 from lxml import html
 from searx.engines.xpath import extract_text, extract_url
-from searx.engines.yahoo import parse_url, _fetch_supported_languages, supported_languages_url
+from searx.engines.yahoo import (
+    parse_url, _fetch_supported_languages, supported_languages_url, language_aliases
+)
 from dateutil import parser
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['news']
@@ -38,7 +41,7 @@ suggestion_xpath = '//div[contains(@class,"VerALSOTRY")]//a'
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
 
-    language = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages, language_aliases).split('-')[0]
 
     params['url'] = search_url.format(offset=offset,
                                       query=urlencode({'p': query}),